Canonical Tensor Decomposition for Knowledge Base Completion
This work addresses knowledge base completion for AI systems, showing incremental improvements over existing methods.
The paper tackled the problem of Knowledge Base Completion by improving Canonical Tensor Decomposition (CP) with a novel regularizer based on tensor nuclear p-norms and a reformulation for invariance to predicate choices, achieving state-of-the-art results on several datasets.
The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) (Hitchcock, 1927) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear $p$-norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.